228 MUTATION AND PLANT BREEDING 



Mi spikes results therefore in elimination of chromosome mutations 

 without reducing the frequency of factor mutations. We have also 

 shown in barley that the M 2 sterility in progenies of partially sterile 

 Mi spikes is actually several times greater than in the progenies of 

 fertile M x spikes. However, in one experiment, even a twofold selec- 

 tion for fertility, namely, in the Mi and M 2 generation, still resulted 

 in the M 3 generation in 3 per cent M 2 spike progenies with partial 

 sterility (34). Usually the breeder is interested in point mutations 

 only and not in chromosomal aberrations. This is valid at least for 

 diploid plants. Selection of fertile Mi spikes means, therefore, an 

 increase of efficiency in breeding with mutations. 



Selection of mutants has commonly been done in the M 2 genera- 

 tion. In that generation only macro-mutations can be recognized. 

 Moreover, owing to the chimeric structure of the Mi plants and their 

 sterility, a great part of mutations are only represented in the hetero- 

 zygous condition in the M 2 generation. These become manifest for 

 the first time in the M 3 generation. Thus, in peas, Gottschalk (37) 

 recognized only 60 per cent of the mutants in the M 2 generation and 

 40 per cent in the M 3 . Similar results have been obtained also in 

 barley (73). If selection of mutants is started in the M 2 generation, it 

 should be continued in the M 3 and following generations. 



The procedure of selection depends on the intention of the 

 breeding program. The selection may be directed towards a special 

 goal, like yield, fungi resistance, baking quality in wheat (58), high 

 protein content in barley (95, 96), germination under low tempera- 

 ture in soybeans (129), strawstiffness, earliness, etc., without any 

 interest for other mutants. Or selection may be nondirected and every 

 mutant which might be of value for plant breeding and which 

 becomes recognized is picked up for further investigations. Of course, 

 directed and nondirected selection may be combined. 



Screening for small mutations can start for the first time in 

 progenies of normal-appearing M 2 plants, and it should be com- 

 menced on a large scale. Selection has to be continued in the follow- 

 ing generations, and the material will become continuously smaller 

 by the elimination of nonmutant and undesired lines. As compared 

 with the pedigree method, for example, the breeder needs a different 

 approach. Breeders using the pedigree method are used to observe 

 large differences among the various progenies and desired types are 



